Results, reality, and recalibration: Unpacking AI hype at SCOPE Summit

I started the month with thousands of other clinical research professionals at SCOPE Summit Orlando. This event always sets the tone for the year – what’s the mood? What are the hot topics? What was big last year, but is now nowhere to be found?

This year, SCOPE was dominated by one big topic: AI. You couldn’t walk five steps without hearing about something that AI was going to fix, transform, or render obsolete. I can’t think of any other topic in my career that has taken up this much conversational real estate. After four days of unpacking the subject with sites and sponsors, here are my biggest takeaways.

The rubber is meeting the road

I heard some say there was too much hype around AI without enough evidence of its impact. I don’t look at it that way; I think we’re witnessing the transition from pure excitement and theory into practice. There are always going to be big, visionary ideas at these events, but we’re also seeing clear traction and results in some areas. I heard from several companies that they’re using AI for interviews to understand patients and design better trials – and that many people are actually more comfortable and candid with AI than they were in traditional focus groups. One research site that we work with told me that she was recently awarded four more studies because of her team’s strong results with AI-powered pre-screening.

One thing I’ll say about separating AI hype from reality: AI doesn’t magically fix broken systems. You can’t layer it into processes that are fundamentally inefficient, disjointed, or exclusive and expect to see transformational results. Take site selection: if sponsors continue to use historical data to select sites study-by-study, they’re going to continue to see the same competition and access challenges they’ve always seen, regardless of whether it is a person or an AI making that decision. But if they take a step back and make a more intentional change to the process – the planning, the timing, the stakeholders, the data – the results can be different.

Sponsors know bigger change is needed, but it’s daunting. The first movers are now taking a hard look at their underlying systems to decide what has to go.

It’s (finally) time for a reckoning on the most tedious trial processes

Some processes have plagued the research industry for decades. They’re inefficient and unpopular, but somehow resistant to change. At SCOPE, the tide finally seemed to be turning.

Feasibility is a good example. This has been a big frustration since I joined the industry in the early 2000s. Sponsors ask the same questions, sites provide the same information, and sponsors (often) turn back to the same sites they’ve used before, just several months later. Now, sponsors are rethinking this. They have mountains of data internally from all the trials that have come before; they’re investing in internal site intelligence platforms that help them make sense of this data.

But one note of caution: relying only on past data from previous trials and familiar sites may fuel familiar problems. If trials are inaccessible or unrepresentative today, AI won’t solve that problem when it’s trained on the same data. Sponsors will have to think carefully about how to keep sites engaged in an evolving process, not rely solely on old insights.

We can’t afford to leave sites behind

Actually, the theme of bringing sites along goes for everything. We’ll definitely never see the full impact of AI if sponsors are the only ones using it, but to date, most of the AI conversation has been about how pharma companies are using AI internally to help teams understand data or sharpen decision-making. Even after SCOPE, there are a lot of unanswered questions about how to incorporate and collaborate with sites in a changing landscape. How will sponsors engage with sites when they move away from questionnaires? How can sites take advantage of AI within the bounds of their contracts with sponsors? What data can sites – especially newer or smaller sites – provide to give them a seat at the table, instead of repeatedly losing trials to huge academic medical centres? These are the types of open questions that need to be addressed head on in the next year.

Change management remains a big challenge, but we can’t let it derail progress entirely

Change management has always been hard for pharma companies; AI has just taken it to a new level. I consistently hear from sponsors that it’s challenging to get new projects launched, and often it’s even more tiring to get everyone else around them on board. There’s real fatigue here, but there are also a couple of clear learnings that we can take forward.

• Change doesn’t happen overnight, but it happens a lot faster when there are clear goals and clear buy-in, from leadership down.
• Start with a clear, known problem; don’t start by searching for something that AI can help with. And don’t try to do too much, too soon. Prioritise one goal or function where improvement is critical and bring people along for the journey.
• Don’t let perfect be the enemy of good. I hear from sponsors all the time that they’re meeting internal resistance to new tech. In site selection, for example, some stakeholders are committed to a handful of sites they already know and trust. My two cents: don’t let partial resistance derail you. If every other site AI helps you choose is a strong performer, you will still succeed.

AI has the potential to make drug development much better, but how quickly and how fully it lives up to this potential depends on what we do next. It depends on the data we use, how we collaborate, and how we empower sites. It depends on our ability to make tough, but necessary, changes to our fundamental processes, instead of sticking AI on as a Band-Aid. It depends on our willingness to take the next step, even when it’s hard, to advance our collective mission of bringing medical breakthroughs to more patients.

About the author

Liz Beatty is co-founder, chief strategy officer, and board member at Inato, where she works towards bringing clinical research to each and every patient. As an advocate for diversity and inclusion, she believes in the opportunity to be more inclusive in clinical trial research to ensure that underrepresented populations are part of trial datasets. Prior to Inato, Beatty headed digital clinical trials at Bristol-Myers Squibb, a pharmaceutical manufacturing company that works to transform patients' lives through science, where she led digital innovation efforts across global clinical operations.

You can listen to Beatty discuss her work in a pharmaphorum podcast here.